Episode 48: Our UK Team Advancing Pharmacogenomics in Healthcare

In Episode 48 of the Informonster Podcast, Charlie Harp is joined by Clinical Architecture’s UK team members Chris Gray and Alex Wren to explore how pharmacogenomics is moving from research into real-world clinical practice.

They dive into the Progress Project, a groundbreaking NHS initiative integrating genetic insights directly into clinical workflows to support safer, more personalized medication prescribing. From building the technical foundation to delivering real-time decision support at the point of care, Chris and Alex share what it takes to turn precision medicine into everyday healthcare.

Tune in to learn how the UK is advancing genomics, improving outcomes, and shaping the future of personalized care.

Contact Clinical Architecture

• Tweet us at @ClinicalArch
• Follow us on LinkedIn and Facebook
• Email us at informonster@clinicalarchitecture.com

Thanks for listening!

Transcript

Charlie Harp (00:01):

Hi, I am Charlie Harp and this is the Informonster podcast. Today on the Informonster Podcast, I'm delighted to have with me two of my colleagues from our office in Exeter in the UK, Chris Gray and Alex Wren. Good morning gentlemen, or good afternoon.

Chris Gray (00:19):

Good morning and good afternoon.

Charlie Harp (00:22):

Hi Charlie. So Clinical Architecture started up the UK office, I want to say has it been 10 years?

Chris Gray (00:33):

Been at least 10 years.

Charlie Harp (00:34):

At least 10 years to service folks in that time corridor and to provide our unique knowledge and capabilities who are friends in the United Kingdom and the NHS and Chris Gray and Alex are Chris Gray. I dunno. I call you Chris Gray. It's like your official name. Chris and Alex. Chris, you're like Madonna. Chris and Alex are instrumental in pretty much all the things we do in the region. And what we're going to talk about today is a project that we've been involved in that we're really excited about. But before we do that, what I'd like to do is have Chris and Alex introduce themselves, talk about what they do and how they got roped into doing this healthcare stuff that we all do know and love. Chris, why don't you kick it off?

Chris Gray (01:21):

Yeah, no, thank you. Thank you Charlie. Really excited to be on one of these. We've listened to lots of these and now to be a part of one is very excited. So yeah, as Joe said, Chris Gray, my current title is Solutions Architect, which spans many roles and areas. But just to give you a brief history of where I started from. I actually started in aerospace working for British Aerospace, implementing large business systems, and I was really kind of, that was my first deployment out of university after a computer science degree of applying technology to what I call tricky business problems. And this was building aerospace equipment, so it was a reasonable tricky side of the business. And then in around the early two thousands, I got introduced to a company called FDB, which really specialized in drug decision support. And I started that and then kind of swapped industries really from aerospace over to health.

(02:25):

Greatly enjoyed the application of really healthcare really, and learning about healthcare from a technology point of view. And worked on a very large project for Boots, the chemist at the time, the pharmacy chain here in the UK. And it was really interpreting a reimbursement set of rules really into a structure and a toolkit, which was a piece of software that would go through and work out how much you'd get for dispensing of medications. And then I joined CA and I've got a broader aspect of went from drug domain across to labs and other kind domains really, and worked on a lot of the shared care initiatives in the UK as well as we call it the pharmacogenomic projects, the PGX projects. So really took a broader side of how we use that data for many applications and the quality of that data. So we've got a much broader role across healthcare, much more levels of interest now, as I said, started aerospace and then moved into healthcare. So that's me, Charlie, and along that journey met both Charlie and Andy Frangleton, who's the current CTO for Clinical Architecture. And I've known Charlie and Andy and Alex for a long period of time now. So thank you.

Charlie Harp (03:49):

You've also been involved in a lot of our international efforts everywhere from China to UAE and appreciate your willingness to get on a plane and go just about anywhere.

Chris Gray (04:03):

Yeah, I did spend a long time in the Middle East and China and I mean, what did you know as you go travel around the world, Charlie, it's the same problem, isn't it? It's just a different flavor. It's tremendously interesting. And I say to Alex now, I couldn't really ever leave this vertical. Really. I love that our small role really can do some good. Really, I love that side of the business really. So that's a very long-winded introduction. So sorry Charlie.

Charlie Harp (04:39):

And what you learn is that every place you go in the world, no matter where you go, they have health data problems, Starbucks and McDonald's. That's comforting and terrifying.

Alex Wren (04:48):

Alright, Alex?

(04:50):

Yeah, thank you Charlie. So my journey into healthcare really started university, studied biomedical science. I liked biology at school, I wanted to study something that had a real application of human biology in practice. So that's kind of how I got into the biomedical side. Straight out of university, I had a brief stint as a lab assistant and that was for a manufacturer of medicated suites. So I mean there's a leap there to some healthcare related stuff, but in that role I was mainly testing for things like omega and menthol levels and vitamin C levels and things like that. I really only mentioned that because similar to Chris, really, I didn't immediately land in healthcare data. And I think I've heard a lot of similar stories in the industry of people that don't immediately find themselves within healthcare data, but once you find it, you kind of stick with it.

(05:49):

So anyway, out of a lab assistant role again, sorry Chris, I'm following you around. I started at FDB in 2011 in the uk and my role there was as a clinical researcher authoring drug clinical decision support focused on medications. And that role really introduced me to healthcare terminologies focused on medications in the uk. At the time that was mainly D, M and D, the dictionary medicines devices. It was CTV three and re two and there was a little thing called SNOMED when I first started. So that role started to give me some insight into what's recorded on your patient record, how it's used within healthcare systems to mean that you get the best treatment that's out there and the clinician doesn't make obvious errors. But also around that time that I was there, there was a big switchover from Reed two to Nommed and that was a big thing.

(06:48):

That was a big deal and I wanted to know why it was a big deal because outside of that world you don't know what those two things are and why they matter. So I started to get an interest into terminologies. That's what really introduced me to informatics and healthcare data and all kind of the good people that work in that world. And there's a lot of them. That introduced me to Clinical Architecture. So I've been able to kind of expand my terminology knowledge outside of the medication world and see what happens internationally as well and get my hands on some clinical data and see how it's used. And that's where I am today. In my role at Clinical Architecture, I'm a healthcare informatics, so I look at different terminologies, how they work, how they fit together in the healthcare system. And yeah, I'd like to think it's all led to being on the Informonster podcast. Charlie, it was all part of the plan.

Charlie Harp (07:43):

Your meteoric rise

Alex Wren (07:45):

I dunno where I go from here.

Chris Gray (07:49):

You could retire.

Charlie Harp (07:52):

I wish. Well what I took away from that is two things, Alex. One is you're probably the three of us. You're the one that's actually qualified to work in health informatics. And the other thing too is you said medicated suites. I knew that the Skittles were curing my head. I knew it. Yeah,

Alex Wren (08:09):

We won't go into e numbers and things on those.

Charlie Harp (08:13):

So the project we're going to talk about that we were involved in this progress project. I think before we get started, it'd be great Alex, since you're the qualified one of the three, if you could give the listeners kind of a primer because what it's about is pharmacogenomics. So if you don't mind, could you give the listeners kind of a primer on what that is?

Alex Wren (08:36):

Sure. And absolutely no pressure. You call me qualified. Yeah. So if you look up the definition of pharmacogenomics, what it will tell you is broadly speaking, it's the study of how a person's genes affects their response to medications. What we'll talk about today really is the application that study in healthcare. So using genetic data to inform prescribing practice for individuals. And if you set the scene of what happens today, the majority of medication prescribing today is based on a one size fits all approach. You present to your clinician with issue X. The clinician goes through their process and they'll prescribe you medication Y. And without sounding too flippant, we'll see how it goes with that medication. You might suffer from some side effects. You might come back and say, this medication's not really working. And that process has been informed by medication trials that say that this thing broadly works for this broad population that we've studied it in at this dose, in this form and in this route.

(09:43):

So we feel safe giving it to you, but we know that everyone's different. Everyone reacts to medications differently. We've all heard stories or perhaps had personal experiences where you have to go back to your clinician after you've been prescribed a medication because it's just not working or you are suffering from side effects. Perhaps you've got a bit of a headache for it and the clinician go, well, there's a second option, we'll give you this. Or perhaps we'll tailor the dose a bit. So sometimes it can be a bit of trial and error. What pharmacogenomics says is that following a pretty simple genomic test, which is often just a cheek 12 or a blood test, I'll get a result that tells me a bit about your genetic record that's obviously specific to you so that when you present with X, next time I can better know or better predict how you are going to respond to the proposed medication that I'm going to give to you for that issue.

(10:42):

So will you perhaps metabolize the medication more rapidly, meaning that you're going to have higher levels of that medication in your body, so you're going to more likely suffer from the side effects? Will you metabolize it slowly so you're not able to reach the therapeutic levels that are required to resolve or help with the issue that you presented? So from that point of view, you start to understand that it's going to be helpful for prescribing in general because you're going to get safer and better medication prescribing so that you don't have to go back to your clinician and tell them that you're suffering with something. What I should also point out there is that the genomic test that you've had then is likely to stay with you for a very long time, if not life, it's very unlikely to change. So the application of that test then becomes valuable for other use cases within the healthcare system.

(11:37):

So in that respect, it's a big hope for the world of medication prescribing a lot less yes, work involved in the prescribing that the medication should make it a lot safer, better outcomes with the healthcare system, fewer visits to your clinicians. There's a of ticks in the boxes when you think about the application pharmacogenomics. I should just point out that pharmacogenomics isn't the only way that people look at the metabolism or the interactions of medications within patients. Obviously other healthcare conditions, other medications, age, all these factors have a part to play. Pharmacogenomics is one strand of that, but you start to get a picture of the impact that this could have on the world of prescribing medications.

Charlie Harp (12:24):

Absolutely, and I think that if you think about it in a single payer system like the NHS, it also reduces spend. You're not spending money on drugs that aren't going to work or are going to produce a negative side effect in a patient. You're reducing the number of times they have to go and engage with a clinician. So not only is it improving the experience and the outcome for the patient, but it's also reducing the potential waste of dollars or pounds on treatments that aren't going to work or are not going to work well. That's

Alex Wren (12:54):

Very cool. Yeah, the economic stack up for this heavily in Its

Charlie Harp (13:02):

Alright. So let's talk about the progress project and when it started, how it started. And I'll leave that to you guys to decide who's going to kick that off.

Chris Gray (13:15):

Yeah, I'm going to start Charlie. So the progress project was really is, I should describe it first of all as actually a clinical trial. Now we think of clinical trials as being new therapies that introduced to the market. This was a clinical trial where patients were recruited to the trial, but they were recruited to really a framework for PGX, A way of delivering that deliver to the point of care. So PGX actually started about three years ago and their first solution was a very, very simple way where the clinicians within let's say primary care in the UK could look up the lab test results. So they'd go onto a website. So that was their first iteration. Then the second iteration on that trial was to actually then say, well instead of the GP having to look up on a website and remember all this information, can we deliver this information naturally the point of care?

(14:20):

So the point of care would be where the clinician was actually prescribing the certain medications and they could get that as what we call a clinical decision alert within the EMR system. And that's really what progress was. And it's meant to be measuring not so much on, there's a lot of information around the benefits of PGX from a research point of view, but it can only be a benefit if it's actually integrated into the point of care. And that's something that hasn't been done before. It was very much a reactive. So if you went in to use Alex's explanation to this, if you went back into the gps, it's not working or I'm experiencing these side effects, they may then order a test. And the only other bit of information the GP might have understood has realized is that genetics could be a contributing factor into this medication that's being prescribed.

(15:25):

But they wouldn't know whether it was a fast metabolizer or a slow metabolizer. This was never really as, I've got my cheat sheet here, it was never really about the science of, as I call PGX really, it was really about the insights into clinical practice. How do we get this system into integrated into the clinician's workflow. So progress is really about building the foundations of the science into everyday practice, which is the application really. This is absolutely the application. So progress was initiated by, so the Northwest Genomic Medicine Service Alliance, so there are seven of these alliances across England, and we're quite lucky in England anyway in that the NHS own these facilities. They even own the labs and the labs systems within those. So they can dictate quite a lot of policy on behalf of the NHS. So the Northwest happens to be a center of excellence for medicines optimization as well.

(16:38):

I hope you don't as well as PGX, as I shorten that word really. And so they were very interested in taking this as I call it a clinical trial, but this kind of foundational trial out to the NHS to build some evidence as the center of excellence that would then, as we take it forward, would then scale across the other regions. And they happen to have, the Northwest is basically Liverpool and Manchester serving seven odd million patients there. And within that they have one of the largest trusts, which is the Manchester Foundation Trust, which is 10 sites across Manchester serving 2.8 million patients serving. We'd also using the EPIC system and we'll talk a little bit about the EPIC systems and the systems we integrated this into. So we set about having really quite a diverse team. We had the Northwest Center of Excellence, we had system suppliers from Primary Care Optum in terms of the uk we probably know that as EISs.

(17:49):

And then we also had a company called TPP. And then we had Epic as another EHR system in secondary care really. So we had the center of excellence, we also had FDB that were providing a medicines optimization solution into those GP systems and they were used as the vehicle to present the alerts at the point of care. So we had lab systems on the sort of, as I got to call it on the left hand side, it's my diagram when I look at it in my head, it's on the left hand side, Charlie. So you've got the lab systems feeding in to then our bit of the project, which was really the foundational architecture for this project. And then on the right hand side we've got the actual GP systems or the Epic system then at the point of care selecting the drug. And then that message coming through to us and we'll go on to talk a little bit about the components on that.

Charlie Harp (18:48):

Excellent. And so for my US listeners, the primary care is acute care or outpatient Secondary care is inpatient, the lift is an elevator and the truck, the car,

Chris Gray (19:02):

Yeah, I knew at some point Charlie, we were going to have this sort of pants and trousers conversation if I can. Yeah, no, so thank you Charlie for that. I couldn't also cope with the interpretation of the language. We will come onto that really because it was quite interesting working with basically geneticists that never really had exposure to clinical systems.

(19:29):

And there I come onto it by say there needs to be a shared language when you start these projects up and it comes off into the lessons learned. So you kind of prompt me there. So let me go onto the second part of that question of what our role is really. Well, I kind of dug this out, but really we were the technical semantic foundation of this solution. So we set right in the center in that there were basically four components and let me go through these components. So we had the lab tests for the patients that signed up to the trial really. And these came out of various limb systems really. And as Alex will probably ate in various formats. So we had to ingest those into what we called the patient store. So we had demographic data as well as genetic data that needed to be processed and stored.

(20:25):

We then had a system that persisted that, all of that structured information and was able to update that information should anything fail in that process. And then we had, as you know, somatic, which acted really as the content management system. And Alex will go on to explain a little bit about that. But there's a mapping really there between the gene, the phenotypes and the drug guidance, which is actually in this case offered by the NHS or the party that were involved in this because they wanted to standardize guidance that would be presented in front of the GP or at the point of care. And that was really important to them to get that interpretation across them. So there was lots of hooks, lots of relationships put into that data. So we could actually then present that information. If you think about it, the GP system is selecting, as Alex described there, a DM and D code that sometime has to make its way right into the information and say, is this DM and D code related to this gene?

(21:33):

Then what is the lab test that's against this gene? And so we are really the glue that brought all that together. And then the last part of that is that we used a FHIR interface really, which is described as CDS hooks. You probably know that in the US marketplace. And that really enabled us, especially when it comes into play with people like Epic who have actually had an interface on a natural interface on CDS hooks. So I think a lot of us would experience with getting changes into or getting interfaces written to large EHR systems. It's not the easiest thing in the world. And I mean we sat in on a hackathon, we spent four hours and we had an integration. So that just shows you the power of aligning to a standards we still had work to do. So they were the four components and I dunno whether Alex you want to, so

Charlie Harp (22:31):

Real quick, can I say that back to you so that see if I understand. So the way it actually operated was you guys were collecting the lab results, keeping it in the central data store repository that correlated the patient identifier with the genetic lab results in that architecture was kind of bound up with the medications that are affected by that gene. The GP system or the system that was interfacing with the implementation was calling CDS hooks passing in a FHIR bundle that had the patient identifier, the drug that was being prescribed as a DM and D code, which is the UK equivalent to RX norm for the US listeners. And if that patient had the genetic test done and it was linked to the DM and D code that was in the FHIR bundle, then it would present back the information, the guidance based upon, hey, if you're going to give this drug, this patient may be a slow metabolizer, yada yada. Is that in a nutshell?

Chris Gray (23:38):

That's exactly right, Alex.

Alex Wren (23:40):

Yeah, absolutely. Easy.

Chris Gray (23:42):

Well done.

Alex Wren (23:42):

Why did we have this big project for it?

Chris Gray (23:46):

Charlie, you've done it in a couple of seconds there. I just said

Charlie Harp (23:50):

It back.

Chris Gray (23:50):

So No, that's exactly it really. And I mean Alex, I dunno whether you want to talk a little bit about the lamp files thing. Well what I will talk about is that what the first part of that component really, and this comes on to the national approach really, is that lab results coming through. Now as part of the trial, we always recognized that that part of the system, the storing of the genetic record, was something that the NHS were very keen to do. And they're very keen and they've got the concept now of just bringing in or trying to bring in the unified genomic record Alex talked about, Alex talked about this being a lifetime record for patients. And the NHS are very keen not to have genetic data only stored in EHRs, which can only be accessed for a particular health facility because it has a lifetime value.

(24:48):

They want it stored in what's called the unified genomic record that is part. So we always realized our one of the four components was only relevant for the trial and the processing of files in the various formats that would come to us was going to disappear in the long time. So for much that what we've built is absolutely relevant and absolutely relevant to the future, but this unified record will kind of take over that patient store and will be a national resource for all users. And I think that's a tremendous benefit to both clinicians and patients really

Alex Wren (25:27):

Just to come in on that point, Chris, a lot of what we've learned about how we process the lab reports that we get and how we store them is totally still relevant to how they're going to be doing this. With a unified record, you've got many different formats. We saw CSVs J files, text files from all of these different labs. How do we ingest or create something programmatically that can understand and ingest and normalize those lab results into a central store? That issue isn't going to go anywhere and you're going to have to have a similar approach to what the project deemed fit for this as well. So there's some themes that will continue on with that. And one thing that I will point out as part of the project that we concentrated on as part of that ingestion was at the point you see that lab result, you've got a great opportunity to validate it.

(26:18):

Is it correct? Is it timely? Is it accurate? Does it contain the right patient ID that you think you should be getting? We've got NHS numbers, they have a check digit, are they right? Do they look okay, have I got duplicate IDs? Is the genetic result that I've got in that lab file, is it actually valid for the gene that it says it's for? So to prevent you from storing all of that potentially data that could go wrong, it's a great opportunity there when you're integrating it into a central shortage to double check it with all of these ideas that you can have to check the data.

Charlie Harp (26:54):

So you're saying there is a potential data quality play there in terms of plausibility alignment?

Alex Wren (26:59):

I think there absolutely should be for somebody storing that level of personalized data to check that it's right and take into account the update process for this. So we say it's a lifetime record, but that might be for a result that's specific to CYP two C 19 for example. Suddenly the research evolves and all of a sudden you can get a genetic result for CYP two C 19 as well as 2D six and all these other genes. But you get retested for them at the same time. So you have a new test that comes in, how do you update that UGR? Do you overwrite what you've already got? Do you check it against what you've already got? So there's opportunity there for the initial population of that to check that it's all valued and nice and neat and tidy. But then when you come to update the record with other genomic results, potentially new variants that are found, how do you know that that's valid as well?

Chris Gray (27:55):

And I think the only other thing I'd say that the size of the prize on this, and it's a messy field. We paint a nice picture Charlie in the ingestion of lab files out, but that's a messy area and it's going to be difficult. But once they do get the information into the UGR, then new research is also coming along at pace around new drugs, new variants that are coming in. So you can still use that data in many different applications and that really is, that probably is the most exciting thing, I think how many, as we call it Alexander, whether you recall, but how many drug gene pairs did we,

Alex Wren (28:38):

You're putting me on the spot. I think there was something like 80 rows of guidance. And what that equates to realistically, that means that one drug might be metabolized by one or more genes and those genes might have different levels around four different levels of metabolization. So you quickly learn that this thing can exponentially grow and grow in terms of the guidance that you can offer for it. I think overall there was something like 26 gene drug as, or gene phenotype drug has,

Charlie Harp (29:15):

And if you think about what's happening right now, the whole genomics field, genomics proteomics is all evolving. And from a terminology perspective, it's a spicy meatball because the data we're talking about is non-trivial. And so really you've got a person's genome, which doesn't necessarily change unless they're doing CRISPR in their garage. You've got the person's genome and then you have what we have discovered or what we believe we've discovered about how certain variants in genes tie to the phenotypes that we're considering and how those things affect the efficacy and contraindications relative to certain medications. And so it's kind of like you think you could stabilize the genetic information relative to a specific human being. And if you have a national identifier you guys have, it's a little bit less challenging than here in the US where we really don't. But that's information that you could stabilize and then you've got that ontology that lives in between that says, if I find this snippet on this gene or this variant of this gene, then that means that these are things I need to be aware of.

(30:35):

And right now we really focus on efficacy and side effects. But there are other things that could be coming down the pike when it comes to treatment modalities outside of medications as we do certain new types of therapies and look at outcomes. The other thing that people don't always consider, I remember years ago I was in Hong Kong and I was representing First Data Bank and they were talking about dose range checking and one of the guys stood up and said, Hey, how come you don't have more dose range checking for neonates? And I'm like, I don't think they do a lot of clinical trials on testing out drug ranges on neonates with all due respect. I don't think I want them to. And it's the same thing with the genomic stuff they're doing. It doesn't just come out of the blue. Somebody does a study that says, we've studied a bunch of people with these genetic markers and we've realized these things indicate whether they're going to be slow, fast metabolizer or whether they're going to have this issue. I think it's definitely, it's an evolving area, but one of the big parts about science is how do we take that science and practically implement it into a workflow so that it matters.

Chris Gray (31:53):

And

Charlie Harp (31:53):

I think what you guys have been doing with the other folks involved in the project in the UK is a great proof of concept of how do we take this evolving science and make it significant to a patient and to a health system. So it's very,

Chris Gray (32:09):

And the general, the clinicians that we worked at, especially in primary care, I think there may have been an expectation that they're experts in genetics, but they're not so part of the whole progress rx. And part of the reason why they offered the guidance is they wanted very clear language to be given in decision support back to the clinician Charlie. So it became really, they were super excited to have, as we both know, we've worked in decision sport long enough to know that we spent a lot of time curating the data and all of those good things. And sometimes it's just switched off. Let's be honest,

Charlie Harp (32:54):

The provider rarely says, thank you for slowing me down. Exactly.

Chris Gray (32:58):

So when we were working in busy GP practices, they were super excited to see the first alert because they knew this was precision medicine. This was something actually, it was more than decision support or maybe I shouldn't say that from legal point of view, but it was information to be trusted and this is why they offered the guidance associated with that. So that I sat back and relaxed and thought, well, this is a different experience. I've not really had a clinician thanked me for alerts in a long while. So it was,

Charlie Harp (33:37):

And Chris, that's the idea of precision medicine and good clinical decision support. 95% of the time the provider should feel glad that the software told 'em about something that maybe they didn't know or they weren't aware of or they didn't catch. And I think the problem we have is in the last couple of decades, we've taken a lot of decision support as an industry and put it in front of a provider and it was very conservative or pessimistic in that it wanted to cover all possible issues. And very often it was not precise, it was very broad and it resulted in providers saying, I already know this, or I'm not worried about this. And it would be very frustrating for them. So once again, this is also an exciting,

Chris Gray (34:28):

Yeah, that's a famous example which is shared by many decision support suppliers is women of childbearing age. Now that is notion that is the opposite of precision. So great. It was a great experience and they were super happy and I think this is a trial, so it will end and the clinicians probably not going to be too pleased about that really. But it all depends on the next steps and how the NHS then going to scale this solution.

Alex Wren (35:06):

I think one point just to mention there as well is that part of the project, there was a clinical team involved in authoring UK specific PGX guidance for this. So adoption then becomes a lot easier for clinicians knowing that and trusting it. It's from the MHS, it's authored in-house IT based on agreed standards in CPIC as well. So the education piece and the alerts for clinicians receiving this advice, it's a lot smoother when that sort of level of authoring can take place.

Charlie Harp (35:38):

So what's the timeline on the trial? When is it supposed to wrap?

Chris Gray (35:43):

Yes, sure. So the new patients, the entrance to the trial closed at the end of December and actually quite recently just asked for an extension for cover up until June six month extension to cover patients to make sure those new patients that joined at the end of December had a runway really. And it closes in June, July of this year. So they'll be looking at the figures and the alerts that were presented in front of the clinicians. They also have a lot of, maybe not so much quantitative research around this particular trial because it wasn't like a clinical trial in that it wasn't saying did it achieve what it was supposed to achieve in terms of therapy. But they've got a lot of connections into the GPS that actually sat in on it really. And they're getting anecdotal what works well, what didn't work well, and then they'll conclude their findings really. But this was quite a lot around what is the process that the NHS needs to go through to get this science into the place where it's needed most, which is completely, I've never been involved in a trial like that in a lot of healthcare. We've been involved in a lot of pilots, Charlie as you know,

Charlie Harp (37:09):

But

Chris Gray (37:09):

We've, we've, they've never tried to look at a system and say, how do we get this to the point of care?

Charlie Harp (37:21):

Any particular lessons learned as you guys went through this? Anything surprise you?

Alex Wren (37:26):

We can do this pretty quickly, Charlie. That's what surprised me. I think we had people from EHR systems, we had ourselves, we have people within the NHS, other industry partners come together pretty quickly. You follow a standard, you know what you're working to, what terminologies you're using. You can implement this pretty quickly and see it working at the coalface and get feedback from users very, very quickly. And I'm fortunate now to have been involved in this, but I haven't seen that speed for a pilot project within the NHS before in my career. So I think it's told me that we can kind of do this pretty quickly.

Charlie Harp (38:07):

That's pretty great. Has there been any pushback on the collection and storage of people's genetic information? I imagine that if we tried to do this here in the states, I'm sure there would be people that would be concerned about that.

Chris Gray (38:21):

No, there was quite a lot of governance that was built in terms of the design around clinical safety and protection of information, all those good things, Charlie. So that is a big part of the project now. And you have to, we weren't, the leaders on this ESMA had to, as part of the trial, had lots of regulations and hoops that they had to jump through really. So I think there wasn't actually, there was no pushback, but that's perhaps because the patient had to consent to the trial. So maybe we see the general acceptance of, I don't want my data accessed inappropriately, but if it's for this type of information and it's well explained and you educate the patient about what or why we're using it, then I don't see that level of pushback. And also, I was going to say with this new being a new project, we considered this as part of the whole steering group at the start, at the very start of the process, it just made governance hoops easiest to jump through.

Charlie Harp (39:31):

Makes sense. And I often wonder, can I tell more about you by your genetic data or your Instagram? I dunno. Yeah, I dunno.

Alex Wren (39:43):

People are feeding their diagnosis into AI already. So you kind of wonder that dichotomy of thing. It exists.

Chris Gray (39:50):

That's right. Yeah, no, the governance's point is an interesting one. We always have this kind of standards debate really. And I do think you need some form of standards to move that quickly,

(40:07):

Especially with the epic example, the CDSX that I spoke about earlier. We'd have said, can we have a request of any modern day EHR in the system? They may say, well, I'll fit you in four years time when I'm on the back end of the customer. But because they had that interface off the bat, we've literally got it working in two to three hours. And we all sat back and thought, well, we had very senior people within the NHS digital world. We sat back and we shared videos and they just couldn't believe it. They couldn't believe it that we'd done a hackathon and we'd got this kind of integrated straight off the bat. An alert was appearing in front of them.

Charlie Harp (40:47):

But let me ask you a question, because that's true when it comes to the DM and D data, that's true when it comes to CDS hooks, but I imagine the whole project would've moved quicker also if there were standards for the genetic data and the lab data and how that data was shared. That sounds like that was probably the biggest chunk of work, was normalizing and rationalizing all the actual lab results that were coming in for the patients, the way you guys describe it.

Alex Wren (41:16):

Absolutely. And I think there's work ongoing about how you can message that lab data within FHIR already. That would just, you'd be almost clicking your fingers and you'd be doing it once you've got that part of the project.

Chris Gray (41:30):

Yeah, had, I mean this a nice example. We had a side angle of this progress project Charlie, which is basically looking at point of care devices, medical devices that you could feed in the sample and it'd give you an output at the point of care. And we fed that all the way through our system and out the other side. And then we did some prescribing on Epic. This is all part of just testing this whole loop really. So the future, and this builds a little bit to the future, Charlie, about devices being in the hospital arena. And Alex, why don't you explain the application, which was MTR one. And

Alex Wren (42:16):

So part of that project was yeah, point of care testing within a hospital that took a sample directly from a patient, processed it, sent that result to a central repository without that sort of need for us coming in, but it could send it directly to the central repository that stores that data and then as soon as that patient is prescribed their medication, that triggers an alert. You've got that immediate guidance presented in front of you. So it's our part of the project that we've looked at, but a bit more cohesive in the fact that that lab result is getting directed straight into that central store.

Charlie Harp (42:54):

That's one of the cool things about a central repository. You can envision a future where based upon my genetic profile here are here's the kind of things that I should eat. Here's the kind of drugs I should avoid, here's the lifestyle choices if I want to stay healthy. You would think you could have this whole map of your, if you want to be healthy and happy, here's the things you should do. And then of course someone will commercialize that and we'll be told to go to McDonald's.

Chris Gray (43:29):

Yeah,

Charlie Harp (43:30):

I'm not jaded, I'm not jaded,

Chris Gray (43:34):

But in the UK they kind of have what's called the integrated care agenda really, which is about me as the patient be following around through the provider system. I might start with mental health, but then I go to physical health, then I go to gp and it's very difficult. We always talk about the NHS, but the NHS is a huge organization with many aspects to it, and it's actually quite difficult for patients to navigate that whole piece. So having a centralized resource that all provider facilities can access is and has to be the model of care.

Charlie Harp (44:14):

Well, and if you think about it, that kind of resource, you think about the concept of whole person care because here in the states we're doing a lot around social determinants of health. Does the person have a place to live? Are they food secure? All those things. Once you have that kind of resource, if you can operate it ethically, you can really do a lot of things to help improve the quality of life for somebody. But we are waxing philosophical at this point, and we could probably do this for hours and hours. So let's do that. Is there anything you guys want to share before we wrap this up? First of all, it's very cool that you've been involved in this and I for one, feel very, I'm very grateful that you guys are part of the Clinical Architecture team. So thank you for all the hard work that you put into this for us. But is there anything you want to share before we wrap this?

Chris Gray (45:14):

Well, I suppose I won't end on a negative, but what next? Really for this really this whole program really, and it is picked up by Center of excellence really, and a lot of the initiatives, this is groundbreaking, but there are some political challenges now and I think we've put our best foot forward. We couldn't have done any better Charlie, in terms of getting this set up and at pace, and we'll wait for the results to come out. But I do, part of unfortunately, what the NHS has always struggled with is how to scale these great ideas. It's a great idea. It needs to be scaled. So part of my pushback would be this is the challenge for the NHS now come on, stop piloting, stop. It was a bit more than a pilot, but now get some benefit.

Charlie Harp (46:12):

It sounds like you've got a good result and they just need to figure out how to roll it out at scale. I mean, take it from a trial to a full blown implementation. Even with continuing to let people to consent. There's a lot of people who I think would be perfectly happy not to feel like they're being experimented on with different medications.

Chris Gray (46:34):

We've got an opportunity as part of the 10 year plan now being resubmitted really. And genomics plays a huge role in what's described as analog to digital really. And like a lot of health economies, we just cannot afford the NHS. It has to get more efficient. It just has to get more modern, otherwise it will fail. And it's already started to cre really. So it's these type of things that are incredibly important. And the 10 year plan picks up just for the UK shared care record, medicine's optimization, personalized care, which the things that we've really talked about today that precision medicine has to be the future rather than women childbearing age clinical decision alerts.

Alex Wren (47:33):

My parting gift is that it can be done. I think that's my big takeaway. As I said before, Chris and I love talking about this project, so if anybody out there is listening and want to know more, then please do reach out to us because yeah, we can talk forever about it.

Charlie Harp (47:48):

I think we should start talking to some of our partners in other parts of the world about what's possible with this and what you guys have seen. I think it's something that in the us, in other parts of the world, I think there's opportunities. So hey, I want to say thank you again, you guys, it has been great. And when we figure out what's going on, we'll have you on again. We can do a retrospective or if you're able to in your amazing careers, do another interesting project. We can talk about that. Alright. All right. Thank you so much. Thank Thanks, Charlie. Chris Gray, Alex Wren, and I am Charlie Harp. And this has been another episode of the Informonster Podcast. Thanks for listening.